1. Technical Field
The present invention relates to a method and apparatus for modeling a room impulse response, and more particularly, to a method and apparatus for modeling a room impulse response that can decrease the number of parameters while maintaining accuracy.
2. Related Art
In general, most of sounds that people daily hear are reverberations, that is, sounds where reflection sounds are mixed. As in a free space, it is impossible to imagine the case where people hear only sounds that are output from sound sources, except for special cases. When people hear sounds in a room, the people feel a space sense, such as the size of an indoor space and the quality of the material of walls, according to a degree of reverberations. In broadcasting and music industries, an attempt to reproduce natural and actual sounds by adding artificially synthesized reverberations to indigenous sounds recorded in a dead room has been made. In recent years, a study has been performed to develop a 3D sound field system to regenerate natural spatial hearing.
A method that is most generally used to obtain a reverberation effect measures a room impulse response in a room that has desired reverberations and performs convolution on the measured room impulse response and indigenous sounds recorded in a dead room. The room impulse response means a sound press that is measured at a location of a listener, when a sound source is excited as an impulse function. In order to model the room impulse response, various methods that include a finite impulse response (FIR)-based all-zero model and an infinite impulse response (IIR)-based pole-zero model are used. In general, a large number of parameters are required to accurately model the room impulse response. In the related art, a sound characteristic of a measured room impulse response is not considered and the entire room impulse response is modeled using one modeling method on a time domain or a frequency domain. When the number of parameters needed to model the room impulse response increases, a data process time and a required memory capacity or processor capacity increase, and a high-performance processor is required. As a result, there occur a lot of restrictions when sound reproduction is processed in real time.